Selective Attraction of Marine Bacterivorous Nematodes to Their Bacterial Food

MARINE ECOLOGY PROGRESS SERIES Vol. 176: 165-178,1999 1 Published January 18 Mar Ecol Prog Ser

Selective attraction of marine bacterivorous nematodes to their bacterial food

Tom ~oens'~',Lesley Verbeeckl, An de Maeyerl, Jean swings2, Magda vincxl

'Department of Biology, Marine Biology Section, and 2Laboratory of Microbiology. University of Gent. K. L. Ledeganckstraat 35, B-9000 Gent, Belgium

ABSTRACT: This paper explores the role of selective attraction to food in determ~ningthe spatial (micro)distribution of closely related nematode species. The attractiveness of 3 different bacterial strains to 4 species of Monhysteridae, Diplolaimelloides rneyli. Diplolaimella dievengatensis, Monhys- tera sp. and Geomonhystera disjuncta, was studied in a multiple choice design. In our study area, the 4 nematode species considered are associated with Spartina anghca detritus decay and have partially overlapping microhabitat preferences. As they all belong to the same feeding guild, they are potential competitors for food. Each of the 4 nematode species was attracted to the bactenal strain B1,but important interspecific differences were noted in the nematodes' response to live or heat-killed bacteria, to bacteria at different cell densities or of different age, and to the filtered supernatant of B1 culture. While the responses of D. meylj to the Gram-positive bacteria Halobacillus trueperi and to the Gram- negative Eschenchia col1 \yere similar, D. dievengatensis and Monhystera sp. were preferentially attracted to H. trueperi and E. coli, respectively. This opposite preference influenced both the numbers and their relative abundances of D, dievengatensis and Monhystera sp. inside bacterial patches in experiments with a mixed 2-species nematode inoculum. Bacterial cell density strongly influenced the nematode response, \nth D. meyliinvariably preferring the highest cell densities offered, while D. dievengatensis and Monhystera sp. had a peak response at lower cell densities. Though chemotaxis IS suggested as an underlying mechanism, the nature of the nematodes' response remains unproved. The present results strongly support the importance of food patchiness in determining the heterogeneous distribution of nematodes, and extend the concept in such a way as to allow for small differences in mlcrohabitat choice between closely related species. They also support the view that nematodes are special~stfeeders, though they probably select spots where suitable food is plentiful rather than individual food particles. Finally, the present study offers a baseline for an understanding and further study of patterns of succession among nematode specles associated with decaying Spartina anglica detritus in terms of highly specific relationships with different strains, growth stages, and densities of bacteria involved in the mineralization of Spartina anglica-derived organic matter.

KEY WORDS: Nematodes . Bacteria . Estuarine . Recruitment . Taxis . Chemotaxis . Microdistribution . Species succession

INTRODUCTION species have been reported for the North Sea alone (Vincx 1989). In the deep sea, diversity may even be The meiofauna of marine and estuarine sediments is considerably higher (Lambshead 1993). By contrast, almost invariably dominated by nematodes. Densities studies trying to streamline this high diversity into in fine-grained intertidal and shallow subtidal sedi- functional groups or trophic guilds have arrived at a ments average 10"nd. m-', representing a biomass of limited number of categories (Wieser 1953, Jensen roughly 0.2 to 2 g C m-' (Heip et al. 1985). An enig- 1987, Moens & Vincx 1997). It has been inferred that matic feature of marine nematode communities is their nematodes are specialist feeders (Tietjen et al. 1970, often high species diversity. It is not uncommon to find Tietjen & Lee 1973, 1977), but observations on selected 50 species in a 10 cm3 core, and, for example, some 800 taxa of the different feeding guilds have indicated a mainly mechanistic food particle selection and significant opportunism (e.g. prey-switching) in the feeding behaviour of several species (Moens & Vincx 1997).

O Inter-Research 1999 Resale of full article not permitted 166 Mar Ecol Prog Ser 176: 165-178, 1999

Meiofauna in general and nematodes in particu1a.r a 1 m2sampling quadrant at the edge of the Paulina salt have a strongly heterogeneous small-scale d.istribu- marsh, situated near the mouth of the Westerschelde tion; the size of their patches may be considerably Estuary, SW Netherlands, where they are mainly asso- smaller than the surface area covered with the tradi- ciated with decaying plant material. They are all tionally deployed 10 cm2 meiofauna cores (Findlay considered deposit feeders (Wieser 1953, Jensen 1987, 1981, 1982). Lee et al. (1977) used a 'cafeteria' setup Moens & Vincx 1997), feeding predominantly on the based on the multiple choice design of Gray (196613) to bacterial flora associated with the plant detritus (Bouw- demonstrate that food patches scattered around a cen- man et al. 1984). A year-round field survey of selected tral meiofauna inoculum attract strongly varying num- microhabitats in the salt marsh (Moens unpubl.) suggests bers of meiofauna and different meiofaunal taxa, significant habitat overlap between the species. As such, depending on the type of food offered. They concluded they are potential competitors for food. that selective recruitment to food spots may be a major In this paper, we focus on differences in the response factor driving the heterogeneous field distribution of of different nematode species (1) to different strains of the meiofauna. In a similar approach, Trotter & Web- bacteria, (2)to bacteria sampled from differently aged ster (1984) demonstrated that 3 dominant nematode cultures, (3) to different densities of bacteria, (4) to bac- species from kelp holdfasts were differentially at- terial growth medium, and (5) to substances released tracted to several types of bacterial and microalgal by the bacteria. The first of these is studied in order to food. The preferences so illustrated revealed a good elucidate the potential of different species of bacte- agreement between the seasonal abundance pattern ria-e.g. those associated with specific types of salt of each nematode species and of its preferred food. marsh detritus-to differentially attract nematodes. Decaying organic matter has been shown to attract The second and third test the hypothesis that nema- some nematode species and, to repel1 others (Buerkel tode species may preferentially respond to bacterial 1901, Gerlach 1977, Riemann 2986, Lorenzen et al. cu.es characteristic of specific stages of detritus decay. 1987, Prein 1988, Olafsson 1992). Gravid females of The fourth and fifth aim at a preliminary characteriza- Metoncholaimus scissus strongly recruited to mycelia tion of the nature of the nematode response. Further- of certain marine fungi (Meyers & Hopper 1966, 1967). more, the hypothesis that any taxis will be influenced In a series of pioneering studies, Gray documented by the abiotic environment is tested using incubations the role of bacteria in determining the horizontal distri- under different temperature regimes. bution of some interstitial archiannelids, a gastrotrich, and a harpacticoid copepod, and demonstrated a highly differential attractiveness among bacteria from the MATERIALS AND METHODS meiofauna's natural habitat (Gray 1966a, b, 1967a, b, 1968, Gray & Johnson 1970). It was concluded that Cultivating the nematodes. A detailed description of bacteri.al films on sand grains differentially attract the methods employed in the isolation, maintenance, meiofaunal organisms, and that the response of the and monospecific, agnotobiotic' cultivation of the meiofauna is mainly directed at characteristics of the nematodes studied is given elsewhere (Moens & Vincx bacterial cell wall, rather than to a product or products 1998). Briefly, spot plates were prepared by the inocu- released by the bacteria into their environment (Gray & lation of small samples of plant litter (Spartina angLica Johnson 1970). Such a response would imply a 'tactile and Fucus vesiculosus) and sediment from the Paulina chemical sense' of the meiofauna as defined by Crisp & salt marsh (Westerschelde Estuary, SW Netherlands) Meadows (1963). Contrary to this interpretation are onto sloppy (0.75%) bacto-agar layers prepared with studies on mainly terrestrial and plant-parasitic nema- modified Killian nutnent medium (von Thun 1966). todes exhibiting a malnly chemotactic response to a h.lonospecific, agnotobiotic cultures of each species variety of inorganic ions, organ.ic .molecules, phero- were established by manual transfer of a few tens of mones, bacteria, and bacteria- or degradation-associ- specimens from the spot plates to a l'% bacto-nutnent ated compounds (see 'Discussion' for references). aga.r (bacto and nutnent agar in a weighwweight ratio The present study aims at elucidating the potential of 4/1) dissolved in artificial seawater (ASW) (Dietnch role of a taxis, i.e. a directed movement, towards patches & Kalle 1957) with a salinity of 25 psu. Bacteria of preferred food in determining the small-scale spatial cotransferred from the spot plates served as food. heterogeneity in nematode abundance and species com- Stocks were kept at 20°C in the dark. By the start of the position. This paper reports on the taxis of 4 monhysterid presently reported experiments, Geornonhystera dis- nematodes towards bacteria. Diplolaimelloides meyli juncta had been in permanent culture for more than Timm, 1966, Diplolaimella dievengatensis Jacobs et al. 6 mo, the other species for more than 1 yr. 1990, Monhystera species Bastian 1865, and Geomon- hysfera disjuncta (Bastian 1865) Jacobs 1987 all occur in 'Containing unidentified associated (micro-)organisms Moens et al.. Neniatode recruitment to bacter~a 167

The 4 nematode species can reach densities of hun- for 24 h at 28°C. Bacteria were then harvested by cen- dreds of individuals per in1 of agar, and as a result of trifugation and subsequently washed 3 times with and the intense microbial activity in the plates, the agar resuspended in physiological water (PW). The super- gradually becomes more fluid. This eventually results natant obtained after the first centrifugation was also in (semi-)liquid cultures dominated by juveniles that used for experiments. do not fully mature anymore, probably as a result of General experimental design and statistical data crowding. When at this stage food is added as a dense analysis. The experimental setup used in this study suspension of Escherjchia coli, growth briefly resumes, was a modified quadrant plate design (Fig. 1)coined resulting in densely populated cultures dominated by from the quadrant plate design of Andrew & Nicholas adults and third (53) and fourth (54) stage juveniles. (1976) and from the cafeteria design of Trotter & Web- Aliquots of such cultures were used in all experiments ster (1984), which in turn are both modifications of with Diplolaimelloides nleyli, Diplolaimella dieven- Gray's (1966b, 1967a, b) multiple choice setup. Candi- gatensis and A4onhystera sp. Geomonhystera disjuncta date attractants and controls were spotted crosswise were hand-picked or rinsed off from the surface of around a central nematode inoculum on sloppy 'bacto- cultures. agar layers. There was always a total of 4 spots sur- Before experiments, nematode aliquots were rounding the inoculum, including the candidate attrac- washed with sucrose in a final concentration of 40% tants and at least 1 control spot. The spots and inocula (w/w) to remove most adhering bacteria and c.ulture were l00 or 200 p1 aliquots of a candidate attractant, medium (Sulston & Brenner 1974, modified according control or nematode culture, with an average distance to pers. comm. of Dr J. Vanfleteren), subsequently between the centers and edges of the nematode inocu- rinsed 4 times in ASW, and finally resuspended in lum and the test spots of 3.5 and 2.5 cm, respectively. ASW. Streptomycin sulphate and benzylpenicillin Sloppy agar layers were prepared by pouring 12 m1 were added in final concentrations of 5000 pg ml- ' and of a 0.5 % bacto-agar (Difco) into 9 cm diameter petn 5000 units ml-', respectively, to block growth of bacte- dishes exposed on a perfectly flat surface. The low ria still present in the nematode inocula. Aliquots from agar concentration enabled the nematodes to easily this nematode suspension were then used for experi- penetrate the agar, as preliminary experiments proved ments. this was important, particularly to Diplolaimella die- Cultivating the bacteria. In each of the experiments vengatensis and Geomonhystera disjuncta. The nema- performed, 1 of the following 4 bacterial cultures were todes were inoculated about 1 h after spotting the can- used: (1) A batch culture isolated from stocks of the didate attractants. The nematode inocula were allowed nematode Diplolaimelloides meyli; this batch culture to evaporate for about 15 min under a laminar flow contained 4 bacterial strains (as determined from hood, because nematodes often were unable to escape observations of colony morphology), 2 of which were the surface tension of the inoculum drop. The petri dominant, grown in 2.5% heart infusion broth dis- dishes were then incubated at 20°C in the dark, except solved in ASM7 with a sallnity of 30 psu (buffered to a when noted otherwise, and the numbers of nematodes pH of 7.5 to 8 with 5 mM Tris-HC1). (2) Strain B1 was in each spot as well as in the sectors between the spots isolated from this batch culture using standard proce- (intersects) counted after 24 h. Preliminary observa- dures and cultured on the same medium. Both batch tions showed that the nematodes dispersed or moved cultures and B1 cultures were grown at room tempera- towards an attractant after seconds or minutes, occa- ture in 250 m1 Erlenmeyer flasks on a rotary shaker. sionally after a few hours, with a stable response For experiments, aliquots of these bacterial cultures almost invariably having been reached after 24 h. were pipetted onto quadrant plates (see 'General Once inside a preferred spot, most nematodes tended experimental design and statistical data analysis'). to stay inside it or make only small excursions in its Alternatively, 'bacteria were harvested from the cul- immediate vicinity (see also Andrew & Nicholas 1976). tures by centrifugation (15 min at 8000 rpm [5200 X g]) A preliminary experiment was run to ascertain that the and subsequently rinsed 3 times with and resuspended position and orientation of the petri dishes inside the in ASW. Th.e supernatant obtained after the first cen- incubator did not influence the dispersal of the nema- trifugation was also used for further tests. todes.

(3) Stock cultures of the Gram-positive bacterium All results have been expressed as relative "(I) of ne- Halobacillus truepen strain BTh4l and (4) the Gram- matodes recovered from spots or intersects. Only the negative Escherichia coli strain LMG2092T were culti- nematode inside candidate attractive spots (includ- vated on 2% marine nutrient agar (Difco). For experi- ing the control spot[s]) were retained for statistical ments, selected colonies of these bacteria were analysis, and the % were adjusted to give composition, inoculated in marine broth (Difco) in 250 m1 Erlen- i.e. their cumulative abundance equals 100%. As such, meyer flasks on a rotary shaker and allowed to grow the weight of all replicates in a replicated statistical test I Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5 I Exp. 6 I Exp. 7

treatment 1

NI medlum

Gd3

treatment 2 treatment 2

Fig. 1. General schematic representation of the modified quadrant plate design used in the present experiments and flow diagram of the setup of each individual experiment. - nematode inoculum, A1 and A2 = candidate attractants, C1 and C2 = control spots. The identity of the different spots in each experiment is given with the following abbreviations: MNI refers to a mixed nematode inoculum, consisting of 73% Diplolaimella dievengatensis and 27% Monhystera sp. B1, Ht, and Ec are the bacterial strains B1, Halobacillus lrueperi BTM1, and Escherichia coli LMG2092T, respectively; Bllm and Bllw are live bacteria in medium and in water (washed),respectively; Bld are heat-killed bacteria; Bllml, Bllm2, and Bllm3 are live B1 in medium from 1, 2, and 3 d old cultures, respectively; B11wl. Bllw2, and Bllw3 similarly refer to live bacteria in water; SN refers to the filtered supernatant of the bacterial culture, and SN1, SN2,and SN3, then, are the supernatant fractions of 1, 2, and 3 d old culture, respectively. 10Xrefers to bacterial densities used. Bxl. Bx2, Bx3, etc. also refer to bacteria at different cell densities, with xl = 10Ycellsml-', x2 = 5 X 10' cells ml-', x3 = 10H cells ml-', etc.. in descending order, following the densities specified in the 'Materials and methods' BC are bacteria from a batch culture consisting of at least 4 different strains. Annotations to the right of the circles indicate species (Dm, Diploma cloides meyli; Dd, Diplolaimella dievengatensis; Msp, Monhystera sp.; Gd, Geomonhystera disjuncla) and number of replicates used per treatment Moens et al.. Nematode recru~tmentto bacteria 169

is equal. Replicated G-tests for goodness of fit to a chi- Expt4: Each petri dish was inoculated with a spot of square distribution, i.e. with 25 % of the nematodes in- washed Halobacillus truepen and one of Eschenchia side each of the 4 spots, were performed in order to de- c011 at opposite sides of the nematode and with 2 con- termine significant deviations from the expected trol spots (Fig. 1) to directly infer any nematode prefer- 1/1/1/1 distribution. Heterogeneity G (G,,) was deter- ence for either bacterial species over the other. Bacter- mined as an indication of whether the observed distrib- ial inocula contained 2 X 108 cells ml-' throughout this utions differed anlong replicates of 1 treatment. Em- experiment. phasis was, however, on the pooled G (G,,),as a measure Expt5: The nematode species with the strongest rel- of overall deviation from the expected distribution over ative preference for Halobacillus truepen, i.e. Diplo- all replicates of 1 treatment (Sokal & Rohlf 1995). Un- lairnella dievengatensis, and the one with the highest planned, pairwise comparisons were performed by relative preference for Escherichia coli, i.e. Monhys- computing GP at a critical probability of a' = a/k, with tera sp., were spotted in a mixed 2-specles inoculum k equal to the number of intended tests (Bonferroni between opposing spots of washed H. trueperi and E. approach, Sokal & Rohlf 1995). Since each interspot col1 at equal cell densities (3 X 10' cells 1111-') (Fig. 1). comparison was potentially meaningful, these un- After 24 h, the total number of each species in each planned tests were performed at an a level of 0.005 bacterial spot was noted, as well as the relative propor- (<0.05/6),ensuring an experiinentwise a of < 0.05. tions of both species in each spot. Response of the 4 monhysterid nematode species to Expt6: The effect of bacterial density on the attrac- the bacterial strain B1. Exptl: Each nematode species tiveness of a spot of Escherichia coli to Diplolaimel- was subjected to 3 different treatments (Fig. 1).In the loides meyli and Monhystera sp. was studied in the fol- first treatment, the 4 spots around the nematode inocu- lowing way. For each nematode species a series of luin were 200 p1 aliquots of (1) a 48 h old B1 culture (= quadrant plates was prepared with spots of decreasing live B1 in medium), (2) supernatant-filtered over a bacterial density. The first plate of such a series had a 0.22 pm millipore filter-of the same B1 culture, (3) spot of log and one of 5 X 10' cells ml-' at opposite sterile heart infusion broth medium, and (4) sterile sides, as well as 2 control spots; the second had ASW of 30 psu. The second treatment consisted of a B1 opposed spots of 5 X 108 and 10' cells ml-l; the series culture spot and an ASW control as in the first series, a was continued down to 103cells ml-' with the following heat-killed (1 h at 70°C) aliquot of the same B1 culture, density pairs: 10' and 107,107 and 106, 106 and 105,and and B1 washed and resuspended in ASW to remove 10' and 103 (Fig. 1). The effect of bacterial density on culture medium. The third treatment had spots of B1 in the attractiveness of H. trueperi to D. ~neyliand D. die- ASW at densities of 101°, 108 and 106 cells ml-l, respec- vengatensis was tested in a parallel experiment. tively, and an ASW control. Impact of temperature on the attraction of nema- Expt2: The attraction of Diplolaimelloides meyli to todes to their bacterial food. Expt7: Aliquots of Diplo- strain B1 at different growth stages was tested using laimelloides meyli and of Diplolaimella dievengatensis 3 treatments. The 4 spots surrounding the central were inoculated amidst 200 p1 spots of live bacteria in nematode inoculum were 100 1.11 aliquots (2 X log cells medium of a 48 h old batch culture, supernatant of this ml-') from cultures grown for 24, 48 and 72 h. In treat- culture, and 2 ASW controls (Fig. 1). Three replicate ment 2, these aliquots consisted of live B1 in medium. petri dishes for each species were incubated in the In treatment 3, they were washed bacteria (= bacteria dark at each of the following temperatures: 5, 10, 15, from that culture but resuspended in ASW), and in 20, and 25°C. After 24 and 48 h, the nematode num- treatment 1 they consisted of culture supernatant. bers inside each spot, in the intersects, and in the ASW was used as the control in all treatments (Fig 1). inoculum spot were counted in order to assess any Attraction to a Gram-positive and a Gram-negative temperature-induced differences in the taxis response bacterium. Expt3: The attraction of Diplolaimelloides and in the activity level of the nematodes. meyli, A4onhystera sp. and Diplolain~elladievengaten- sis to the bacteria Halobacillus trueperi and Escher- ichia coli was studied. Geomonhystera disjuncta was RESULTS omitted from this and all subsequent experiments, as well as from treatment 2 in Expt. l, because of an Response of the 4 monhysterid nematode species to infection of the stock cultures. For each nematode the bacterial strain B1 species, quadrant plates were spotted with H. trueperi in PW (= washed bacteria) and with the filtered culture Expt l supernatant at opposite sides, and with 2 spots of PW as controls, and incubated at 20°C in the dark for 24 h. The results of the experiments on the attraction of A second treatment was run with E, c011 (Fig. 1). the 4 nematode species to the bacterial strain B1 are 170 Mar EcoI Prag Ser 176: 165-178.1999

Table 1.Relative recruitment percentages of the 4 monhysterid nematode species to the unidentified bacteria1 strain Bl.Aver- ages and standard deviations of 2 or 3 replicates per treatment are given. nd = not determined. Washed bacteria are live culture aliquots washed and resuspended in ASW (see 'Materials and methods']. All bacterial spots had cell densities of 10" cells d.', except for bacteria 1/100and bacteria 1/10000,which had densities of 10' and 106 cells ml-', respectively

Treatment 1 Live Bacterial Supernatant Control Total Central bacteria medium intersects inoculum

Monhystera sp. 41.5* 2.12 0 * 0 5 * 4.24 13 * 2.83 21.29 * 4.39 t9.42* 0.14 Diplolaimelloides meyli 37 * 2.83 9.5* 4.95 16 + 4.24 9.5* 0.71 23.77 * 2.14 4.6 * 0.97 Diplolaimella dievenga tensis 12.5 0.71 1 * 1.41 12 * 1.41 17 * 4.24 17.69 * 10.08 36.45 * 14.33 Geomonhystera disjuncta 27.2 * 8.3 1.5* 1.7 7.4 *3 14 * 8.2 7.6 * 4.7 39.3 * 10.5

Treatment 2 Live Heat-killed Washed Control Total Central bacteria bacteria bacteria intersects inoculum

Monhystera sp. 10 i 4.24 3 * 2.83 27 i 2.83 7.5 i 3.54 8.31* 0.18 43.81* 2.13 Diplolaimelloides meyli 23.5* 0.71 35.5* 2.12 20 * 2.83 5*0 13.06 i 1.68 3.49 * 1.78 Djplolaimella dievenga tensis 13.5* 16.26 O*O 7.5 * 17.68 4.5* 3.54 11.451t8.41 42.14* 14.19 Geomonhystera disjuncta n d n d nd nd nd nd

Treatmen: 3 Washed Bacteria Bacteria Control Total Central bacteria 1/100 1/10000 intersects inoculum

Monhystera sp. 37.5* 12.02 13 +- 9.9 10.5 i 2.12 4.5 * 0.71 3.35* 1.28 31.59 c 7.06 Diplolaimelloides meyli 33i11.31 23k424 14.5*4.95 12 5 5.66 10.28+ 0.02 5.58* 2.51 Diplolaunella dievengatensis 14.5 * 4.95 15.5i 7.78 24.51: 7.78 8.5 + 7.78 7.13* 0 65 29.8+ 1.74 Geomonhystera disjuncta 33.5* 13.03 13.7* 7.27 9.1+ 1.94 7.1* 2 05 13.6 + 3.91 23 + 7.11

summanzed in Table 1. In the first treatment, live bac- Expt 2 teria in medium attracted significantly higher numbers of Monhystera sp., Diplolaimelloides meyli and Live B1 in medium from 3 d old culture attracted Geomonhystera disjuncta, but not of Diplolaimella die- more Diplolaimelloides meyli than did bacteria from vengatensis, than did control spots (p G 0.001). D. 2 d old cultures (p < 0.005), but differences between 1 meyli was the single species not to be repelled by bac- and 2 d old or between 1 and 3 d old culture aliquots terial growth medium and to be attracted to super- were not significant (Fig. 2). No differences were natant of bacterial culture. The attractiveness of the supernatant was lost when diluted with an equal vol- ume of ASW, or after heating (1 h at 60°C) or autoclav- ing (l5 min. at 1.1 atm and 120°C) (data not shown). In 45 the second treatment, washed bacteria were more 40 attractive to Monhystera sp. than unwashed culture $ 35 * -- aliquots (p <. 0.001), while both attracted similar num- 30 bers of D. meyli. The latter species, however, signifi- E E 25 cantly preferred heat-killed over live ba.cterid (p < 5 0.001). The response in D. dievengatensis was highly 20 E heterogeneous among replicates (p @ 0.001), but this 6 ,, species was repelled by heat-killed bacterial cells in S 10 both replicates (p < 0.001). In the third treatment, the 5 highest cell density attracted significantly more nema- 0 todes than did lower densities in all nematodes except B1 in washed B1 supernatant D. dievengatensis. The response was density depen- medium dent over the entire range of observed densities in D. meyli (p 0,001), while the response of ~~~h~~~~~~ Fig. 2. Diplolaimelloides rneyli. Effect of bacterial culture age on the attractiveness of culture aliquots, of washed bacteria, SP. and G. djsjuncta to the lower 2 densities did not dif- and of filtered culture supernatant to the nematode. Means fer (p > 0.05). D. djevengatensis preferred the lowest and standard deviations of 3 replicates per treatment are cell density over the 2 higher ones (p < 0.005). given Moens et al.. Nematode recruitment to bacteria 171

H. trueperi Expt 4

When simultaneously presented with the 2 species of bacteria, Diplolaimelloides meyli was equally attracted to both; Diplolaimella dievengatensis significantly (p 4 0.001) preferred Halobacillus trueperi over Escherichia coli, while Monhystera sp, exhibited the opposite preference (p < 0.001) (Fig. 4). In the former 2 species, there was significant heterogeneity among replicates (p < 0.005). Pooled data suggested a preference of D. meyli for H. truepen over E. coli, but omission of the deviant replicate from the analysis overruled this effect. Replicate heterogeneity did not E coli affect the observed preference of D. dievengatensis for H. truepen.

Expt 5

In a mixed inoculum of Diplolaimella dievengatensis and Monhystera sp. simultaneously offered Halobacillus trueperi and Escherichia coli, similar numbers of Monhys- tera sp. were found in both bacterial spots. D. dievengatensjs, on the other hand, was 3 times more abundant in spots of H. trueperi than in Fig. 3. Diplolairnelloides meyli, Monhystera sp. and Diplolaimella die- spots of E. coli (Fig. 5). As a consequence, the vengatensis. Relative recruitment percentages of the 3 monhysterid relative percentages of D. dievengatensis and nematode species to cells and filtered supernatant of cultures of the bacteria Halobacillus truepen and Escherichia coli. Means and stan- Monhystera sp., which in the inoculum were dard deviations of 4 replicates per treatment are given. See 'Materials 73 and 27, respectively, decreased to 62.8 for and methods' for details on treatments, incubation conditions and D. dievengatensis in E. coli spots and to 22.8 bacterial cell densities used found between the numbers of D. meyli reaching 0 Monhyacm sp. supernatant of 1, 2 or 3 d old cultures. However, washed bacteria from 1 d old culture attracted twice as T many D. meyli than did bacteria from 2 and 3 d old cultures (p 0.001) (Fig. 2).

Attraction towards a Gram-positive and a Gram- negative bacterium

Expt 3

Diplolaimelloides meyli moved to both washed Halobacillus trueperi and Eschenchia coli, but not to Fig. 4. Diplolairnelloides meyli, Monhystera sp. and D~plo- their respective supernatant fractions (Fig. 3). D. die- laimella djevengatensis. Relative recruitment percentages of vengatensis showed a small but significant (p < 0.001) the 3 monhysterid nematode species to the bacteria Halo- bacillus trueperi and Escherichia coli offered simultaneously. positive response to both cells and supernatant of H. Data are means and standard deviations of 4 replicates per trueperi but not of E. coli. By contrast, Monhystera sp. treatment. See 'Materials and methods' for details on treat- reacted only to E. coli cells. ments, incubation conditions and bacterial cell densities used 172 Mar Ecol Prog Ser 176: 165-178, l999

highest cell density (10' cells ml-l) over lower ones (Flg. 6). The response was, however, not entirely density depen.dent over the whole range of densities tested, with a fairly density-independent attraction in the intervals of 108 to 5 X 10' cells ml-' and 103 to 107 cells ml-'. A sharp decline at densities below 108 cells ml-' was obvious. Diplo- laimella dievengatensis, however, preferred a cell density of 5 X 108 ml-' over the higher (10') and lower ones. The rest of its response was broadly similar to that of D. meyli, with a fairly density-independent response in moculum H. mcperi E coli similar intervals, and with a steep decline at the transition from 108 to 107 cells ml-' (Fig. 6). By contrast, Monhystera sp. showed a density-dependent response over the

300 10Monhystera sp. 1 whole interval tested, with peak numbers - reaching spots of 107 Escherichia col1 ml-'; 250 higher and lower bacterial densities re-

200 cruited less Monhystera sp. (Fig. 6).

I 50 Impact of temperature on the attraction of 100 nematodes to their bacterial food 50 Expt 7 0 mo~ulurn FI.Initpen E col1 After 24 h, the respective percentages of nematodes that had moved out of the inocu- Fig. 5. Djplolaimella dievengatensis and Monhystera sp. (A)Relative densities and [B) absolute numbers of the nematodes inslde spots of the bac- lum at 5l 151 20f and 250C were teria HalobaciLIus trueped and Escherichia coli in incubations with mixed approximately 2, 20, 20, 50, and 70% for inocula of both nematodes between 2 bacterial spots at oppos~tesldes. Djplolaimelloides meylj and only 0, 4, 12, 12, Means and standard deviations of replicates are shown 3 and 12 % for Diplolaimella dievengatensis (data not shown). In the latter species, however, the inocula had not evaporated and for Monhystera sp. in spots of H. trueperi, and many individuals were at the edges of the drops, increased to 77.2 for D. dievengatensis in spots of H. unable to escape the surface tension. After 48 h, the truepen and to 37.2 for Monhystera sp. in spots of respective percentages were approximately 5, 65, 75, E. coli. The differences between relative nematode 90, and more than 95'"0 for D. meyli, and less than 1, 5, abundances in bacterial spots were significant (p 10, 40, and 70% for D. dievengatensis (Fig. 7). A paral- 0.005), as were the relative depletion of D. dieven- lel response was noted for the numbers of D. meyli gatensis and the relative enrichment of Monhystera sp. inside the bacterial spots after 24 h; after 48 h, how- in the E, col1 spots compared to the inoculum (p < ever, half of the nematodes had reached the bacterial 0.001). The percentage increase and decrease of D. spots at all temperatures from 10 to 20°C. At 5 and dievengatensis and Monhystera sp., respectively, in 25"C, this was less than 5 and more than 75 %, respec- the H. trueperi spots relative to the inoculum, though tively. In D. dievengatensis, less than 10% of the occurring in all 3 replicates, was not statistically signif- nematodes had reached the bacterial spots at tempera- icant (p > 0.05). tures up to 15OC, even after 48 h, and the bactenal spots were not significantly more attractive than were the controls. At 20 and 25"C, however, approximately Expt 6 35 and 60% of the nematodes were inside the bacterial spots after 48 h., while the controls at these tempera- In trials with different densities of bacterial cells, tures contained equal numbers of nematodes as those Diplolaimelloides meyLi con.sistently preferred the at the lower temperatures. Moens et al.: Nematode recruitment to bacteria 173

90 In a first experiment, we show that the +Monhystera sp 80 bacterial strain B1 is attractive to all 4

70 monhysterid nematodes tested, but that the conditions under which this bac- 60 terium attracts nematodes largely differ. 50 This highly differential response sug- 40 gests that even 1 single bacterial species 30 might be able to influence a commu- 20 nity of the 4 nematodes studied to form

10 patches of at least 3 different relative species compositions, depending on the 0 111 1R 115 1110 11100 lllOd 11106 'condition' of the bacterial food. The bacterial concentration (111 = 10~cellsper rnl) observation that Diplolain~elloides meyli res~ondeddifferentlv to bacteria sam- Fig. 6. Diplolaimclloides meyli, A4onhystera sp. and Diplolaimella dieven- ple> from cultures of different age fur- gatensis. Influence of cell density on the recruitment of the 3 nematode species to bacterial spots. Recruitment at the highest cell density was normalised ther supports the potential of bacteria in to 50%. In reality, it averaged 57, 24, and 24% in D. meyli, Monhystera sp., different phases of growth to differen- and D. dievengatensis, respectively. Recruitment percentages at the lower tially attract nematodes. Cells from cul- cell densities were extrapolated by comparison of data at a particular cell tures in exponential growth phase are density to the mean recruitment at the previous density, with densities preferred over cells from older cultures ranked from highest to lowest. Each data point represents the mean of 3 replicates. Error bars show the varlance as a percentage of the mean. S~ncein D. mey'i. Evidence for an impact of all but the h:ighest and lowest cell densities, 2 separate series of 3 replicates bacterial (nutritional) status on the each were counted, the highest of both variances is shown migration of nematodes was also pre-

DISCUSSION D. dievengatensis A considerable body of literature exists on the response of mainly plant-parasitic and terrestrial nematodes to a variety of external stimuli, including electrical, mechanical and chemical stimuli and factors such as temperature and light (reviews in Croll 1970a, Dusenbery 1980, Coomans & De Grisse 1981, Huettel 1986, Perry 1996). Analogous information on marine or brackish-water nematodes is, however, scant, and limited to observations on the attractiveness of CO2 to Adoncholajmus thalassophygas (Riemann & Schrage 1988) and the recruitment of different major meio- fauna1 taxa or different nematode species, belonging to different feeding types, to patches of candidate food or towards sediment impregnated with different species of bacteria or unicellular algae (see 'Introduction' for references). This study demonstrates a highly species-specific marine nematode response to food and shows that one candidate food organism may be, depending on its condition, attractive, unattractive or even repulsive to the same nematode species.

0 5 10 15 20 25 Fig. 7. Diplolaimella dievengatensis and Diplolaimelloides temperature (OC) meyli. Recruitment response of nematodes to inocula of the -- - bacteria Halobacillus truepen at different temperatures and +bacteria +total intersects after a 48 h incubation. Data are averages and standard -m- control +centre deviations of 3 replicates 174 Mar EcolProg Ser 176: 165-178, 1999

sented for the terrestrial Caenorhabditis elegans blages was corroborated by our results: D. dieven- (Grewal & Wright 1992). gatensis increased its dominance over Monhystera sp. In a second step, we focused on the response of ne- in the H. trueperi spots, but decreased relative to matodes to different species of bacteria. Three different Monhystera sp. in E. coli spots. In addition, the total types of response were noted for the 3 Monhysteridae numbers of nematodes reaching food spots heavily studied. no clear preference (Diplolaimelloides meyli), depended on the nature of the bacteria, with H. a preference for the Gram-negative bacterial strain trueperi invariably attracting far higher nematode over the Gram-positive one (Monhystera sp.), and the numbers than E, coli. While previous studies have reverse (Diplolaimella dievengatensis). In summary, already reported on the differential potency of a num- Halobacillus trueperi was attractive to all 3 Monhysteri- ber of bacterial strains to attract Caenorhabditis ele- dae, but was the preferred source for only 1; Escher- gans (see above), the present results demonstrate that ichia coli elicited a positive response from only 2 nema- 1 bacterial strain can at the same time elicit a positive tode species, and was the preferred source for 1 of response from one nematode, but not from another, these. A differential attra.ctiveness of different bacterial closely related species. The observed preferences species to terrestrial nematodes (Andrew & Nicho1.a~ are unlikely to relate to preferences for either Gram- 1976, Jansson & Nordbring-Hertz 1983, Grewal & positive or Gram-negative bacteria, since, for example Wright 1992), and to free-living stages of the insect- D. dievengatensis, the species with the most pro- parasitic Neoaplectana carpocapsae (Pye & Burman nounced preference for H. trueperi, has successfully 1981) has been noted previously. Caenorhabditis ele- been cultivated on diets of Gram-negative bacteria gans showed either a strong, an intermediate or a weak (Vranken et al. 1984). positive response to different bacteria and was repelled Gray (1968) found no significant influence of bacter- by others (Andrew & Nicholas 1976, Grewal & Wright ial density on the attractiveness of bacteria to the 1992). However, none of the cited studies tested for the harpacticoid Leptastacus constrictus, while Gray & degree of response or preference in multiple choice Johnson (1970) did note a significant correlation experiments with 2 or more strains offered simultane- between the number of attractive bacteria and the ously, nor were bacteria standardized to cell densities. response of the gastrotrich Turbanella hyalina. The dif- This study therefore eliminates the possibility that the ferential density-dependent response of the 3 nema- observed preferences might be due to a density-depen- todes used in our experiments to bacteria offers an dent response. However, there still remains a chance attractive basis for explaining observed microhabitat that the choice of a non-specific growth medium for preferences in terms of a succession of species on the 2 bactenal strains could have affected the (nutri- detritus in different stages of decay. As such, Diplo- tional) quality of the bacterial populations in our exper- laimelloides meyli could somewhat presumptuously be iments, and as such influenced their attractiveness to considered as the species that may most readily nematodes. respond to the early breakdown of plant litter, at a Surprisingly, in tests with a mixed nematode inocu- stage where concentrations of highly labile organic lum consisting of Diplolaimella dievengatensis and carbon and correspondingly high densities of bacteria Monhystera sp., the latter species did not exhibit the are available. Diplolaimella dievengatensis and Mon- same preference for Escherichia coli over Halobacillus hystera sp. might then be envisaged as preferentially truepen that was observed in monospecific nematode associated with later stages of leaf litter decay, i.e. with inocula. D. dievengatensis, on the other hand, did generally more refractory material and lower overall show the same pattern of response. We repeated this bacterial densities. Although 0u.r results provide no experiment with observations of the nematodes' direct evi.dence for this relation, it is noteworthy that migration after shorter incubations, and found that D. in situ, D. rneyli is most abundant on decaying leaves dievengatensis responded more rapidly to its preferred still attached to the stems of Spartina anglica and other source (i.e. H. trueperi), while Monhystera sp. started macrophytes, whereas D. dievengatensis and Mon- migrating somewhat later. Initially, this migration was hystera sp. are more typical of the sediment around the mainly directed versus the E. coli inoculum, but many roots of macrophytes, where they could be associated individuals reversed before reaching this spot and with the burial of plant litter. Geornonhystera disjuncta started migrating in the opposite direction. We suggest takes a position that is more similar to that of D. meyli, that the tracks of D. dievengatensis in some way in- but prevails at lower temperatures (Moens unpubl.). fluenced Monhystera sp., setting out a pattern that It is also interesting to note that in rnonoxenic cultures directed a larger-than-expected fraction of this spe- of G. disjuncta and D. dievengatensis on the bacter- cies' inoculum to the H. trueperi spot. Inspite of this, ial strain Alterornonas haloplanktis IS&, the former the hypothesis that different bacterial spots would be nematode needed high food levels, whereas the latter colonised by differently composed nematode assem- (erroneously referred to as Monhystera microphthalma Moens et al.: Nematode r,ecruitment to bacteria 175

in the original article) thrived on cell densities as low tors emanating from prey or host organisms govern the as 106 to 10' bacteria ml-' (Vranken et al. 1984). Fur- primary food-finding mechanisms in nematodes (Croll thermore, the density-dependent response also indi- & Sukhdeo 1981, Zuckerman & Jansson 1984); both cates that the mere presence of a bacterial cue may be 'taste' (most studies) and 'smell' (Bargmann et al. 1993) insufficient to trigger a nematode response: not only are involved in this chemotaxis. From the present may the concentration of the stimulant be too low, it observations on the nematodes' migration up bacterial may also be too high. cue gradients, it is likely that the recruitment in our It is by no means surprising to find that both the experiments also resulted from a chemotactic re- overall motility and the taxis of nematodes towards sponse, and, since the time allowed for the establish- bacteria are strongly dependent on temperature. In ment of gradients emanating from the attractive spots general, the activity pattern of both species studied was relatively short (sometimes less than 1 h) and the agrees well with patterns of temperature dependence nematodes' response often instantaneous, both soluble as established from life cycle studies (Vranken 1985, and volatile substances may have been involved. for Diplolaimella dievengatensis; Moens unpubl. for Nevertheless, the nature of the stimulus or stimuli Diplolaimelloides meyli) and from measurements of that guide bacteriophagous nematodes to their food respiration (Moens unpubl.). All these data point at is hitherto unknown. The only bacteria-associated temperature optima in between 20 and 30°C for both semiochemical of which the involvement in nematode species, with D. dievengatensis perhaps preferring chemotaxis has repeatedly been demonstrated is CO2 slightly more elevated temperatures than D, meyli. (Klingler 1965, Edmunds & Mai 1967, Croll 1970a, Whereas both nematodes appear to have a similar tem- Dusenbery 1974, Pline & Dusenbery 1987, Riemann & perature optimum, their activity at lower temperatures Schrage 1988). The attraction of free-living nematodes (up to 15'C) as revealed by their migration away from to a variety of inorganic ions (Ward 1973, Dusenbery the inoculum spot differs, with D. meyli remaining 1974, 1976a), cyclic nucleotides (Ward 1973), and other more motile than D, dievengatensis at 10 to 15°C. organic compounds (Dusenbery 1975, 1976b) has been However, since similar numbers of nematodes were documented. A remarkable specificity of the response recovered from control spots and intersects at all tem- has been noted, e.g. D-tryptophan repelled Caenor- peratures after a 48 h incubation, the distinctly higher habditis elegans (Dusenbery 1975) and electrophysio- numbers and proportions of nematodes inside bac- logically stimulated the parasitic Syngamus trachea terial spots at the higher temperatures suggest that the (Riga et al. 1995), while L-tryptophan elicited no re- efficacy with which both nematodes respond to the sponse from C. elegans (Dusenbery 1975). bacterial spots is also temperature dependent. This can Although observations of the nematodes' behaviour be explained either by a better perception by the in our experiments are highly suggestive of a true taxis nematodes of the bacterial stimulus, or alternatively by response, the possibility of a random food-finding an increased stimulus production by the bacteria in strategy where nematodes stay inside a suitable food the plates. spot after a chance encounter cannot be entirely This paper demonstrates that free-living marine excluded on the basis of this type of experiment alone. nematodes migrate in a directed way towards patches In a heterogeneous environment such as the benthos, a of food. In the absence of an attractive source, nema- random food-finding strategy would be disadvanta- todes showed a random movement on and in the agar. geous, particularly in organisms with an overall low Deviations from this 'random walk' behaviour con- vagility (White 1978). Even in organically enriched sisted of clustering in groups of several tens of indi- sediments of the Westerschelde Estuary, on average viduals in the inoculum spot or migration of adult less than 3 % of the sediment consists of organic mat- males to female 54 and adults. Clustering of nematodes ter, of which only part is a potential food source for the in aqueous suspensions has been noted previously and meiofauna. The difficulties in monoxenically rearing has been ascribed mainly to mechanistic interactions marine bacteriophagous or herbivorous species (see among individuals in dense suspensions (Doncaster & Moens & Vincx 1998 for a review) are evidence of Webster 1968, Croll 1970b); in our experiments, Diplo- highly specific nutritional requirements, which are met laimella dievengatensis showed the strongest ten- by only a few food organisms, and the efficient finding dency to aggregate, probably because the D. dieven- and recognition of these suitable foods can therefore gatensis inocula contained on average the highest be considered vital to the nematodes' reproductive nematode densities. Consequently, the overall re- success. sponse in D. dievengatensis was less than in the other The assay method used in our experiments is a rapid nematodes studied. and suitable method for primary assessment of the There is a general consensus, based on morphologi- response of many marine and brackish-water nema- cal and experimental evidence, that chemotactic fac- todes, be they cultivated or extracted from sediment 176 Mar Ecol Prog Ser 176: 165-178, 1999

samples, to a variety of candidate food sources. It is, ature patterns for their positioning in soils (Dusenbery however, much less applicable when aiming at the 1988, 1989, 1996, Robinson 1994). Plant parasites may identification of particular single stimuli involved in orientate along redox-induced gradients (Bird 1959) or the nematode response. For that purpose, individual electric fields (Robertson & Forrest 1989) around plant nematodes need to be studied in a proper gradient of roots. They may respond to subtle changes in pH the stimulus. Candidate stimuli can preferentially be (Ward 1973), and structural heterogeneity, which is administered to the center of a petri dish containing a omnipresent in a benthic environment, has recently homogeneous agar or sephadex layer, and after appro- been shown to interact with chemotaxis to give com- priate time has been allowed for the establishment of a plex patterns of attraction to bacterial cues (Anderson radial gradient, single nematodes can be inoculated to et al. 1997a, b). There can be little doubt that taxis is the edges of the dish and their response noted (Ward not restricted to nematode-bacteria interactions, but 1973, Riga & Webster 1992). Additional information on equally mediates the response of nematodes with other the nematodes' response can be obtained from de- feeding strategies. Attraction of nematodes to exudates tailed observations of their migration, e.g. by photo- from yeasts and fungi (Balanova et al. 1979) and re- graphing nematode tracks (Riddle & Bird 1985, Riga & pulsion from slime molds (Kessin et al. 1996) have Webster 1992) or by video-monitoring of their move- been demonstrated. The principle illustrated here may ment (Dusenbery 1983, 1992, Pline & Dusenbery 1987, therefore generally govern nematode-food interrela- Anderson et al. 1997a). Totally different approaches tions in the benthic environment, and as such deserves towards the study of nematode responses to chemicals further study. have been countercurrent separation (Dusenbery 1973, 1974) and electrophysiological stimulation of tethered Acknowledgements. The first author performed most of the worms (Riga et al. 1995, Perry 1996). presently reported work under the auspices of a grant as aspi- The present results reconcile the seemingly contro- rant with the Fund for Scientific Research-Flanders (FWO). versial observations of a high selectivity (Tietjen et al. Further financial support was received from the Flemish 1970, Tietjen & Lee 1973, 1977) and a mainly mecha- Ministry of Education via GOA contract 92/98-08 to the Uni- nistic and rather unselective food ingestion (Moens & versity of Gent, from the Belgian Federal Ministry of Science via the Impuls Programme Sea (MS/02/080), and from the Vincx 1997) in the feeding of marine nematodes. We University of Gent via BOF 98-03, contract no. 12050398. suggest that many nematode species select spots Johan Vandenberghe and Cindy Snauwaert skillfully assisted where suitable food is abundant from a distance, but with the bacterial culture work. Three anonymous referees may feed rather non-selectively therein. Their re- gave constructive comments on an earlier draft of the manu- script. Drs Peter Herman, Jarle Tufto and Konjev Desender sponse to food is thus highly selective, but their inges- gave valuable comments concerning different possibilities for tion may be less so. the statistical analysis of our data. Chemotaxis is considered an important factor underlying the patchiness of benthic harpacticoid copepods, LITERATURE CITED though passive dispersal through hydrodynamic forces may be of equal importance (Fleeger et al. 1995, and Alkemade R, Wielemaker A, Herman PMJ, Hemminga MA references therein). The nematodes studied here are (1994) Population dynamics of Diplolairnelloldes bruciel, a typical 'Aufwuchs' species, and probably subject to nematode associated with the salt marsh plant Spartina considerable hydrodynamic d.isturbance, as most of anglica. Mar Ecol Prog Ser 105:277-284 Anderson ARA. Young IM, Sleeman BD, Griffiths BS, Robert- their habitat is inundated at high tide. In a similar habi- son WM (1997a) Nematode movement along a chemical tat, the densities of Diplolaimelloides bruciei were sig- gradient in a structurally heterogeneous environment. 1 nificantly reduced by flooding (Alkemade et al. 1994). Experiment. Fundam Appl Nematol20:157-163 The capacity to efficiently find suitable feeding spots Anderson ARA, Sleeman BD, Young IM, Griffiths BS (1997b) Nematode movement along a chemical gradient In a may therefore be vital to nematodes living in these structurally heterogeneous environment. 2. Theory. Fun- dynamic environments. It is as yet unclear at what dam Appl Nematol20: 165- 172 distances food spots can be recognized and how the Andrew PA, Nicholas WL (1976) Effect of bacteria on disper- nematodes move towards them. Jensen (1981) demon- sal of Caenorhabditis elegans (Rhabditidae). Nematolog- strated that, next to movement on or through a sub- ica 22:451-461 Balanova J, Balan J, Krickova L, Nemec P, Bobok D (1979) strate, some nematodes may show a chemotactically Attraction of nematodes to metabolites of yeasts and fungi. driven swimming behaviour. J Chem Ecol5:909-918 Food is clearly not the only factor underlying the Bargmann Cl, Hartwieg E, Horvitz HR (19931 Odorant selec- patchiness of benthic and Aufwuchs environments. tive genes and neurons mediate olfaction in C elegans. Cell 74:515-527 Nematodes have been shown to respond with almost Bird AF (1959) The attractiveness of roots to the plant para- incredible accuracy to temperature gradients (Hedge- sitic nematodes Meloidogyne javanica and M. hapla cock & Russell 1975, Thomas 1995) and to use temper- Nematologica 4:322-335 Moens et al.. Nematode I.ecruitment to bacteria 177

Bouwman LA, Romeyn K, Kremer DR, van Es FB (1984) cesses associated with the initial settlement and post- Occurrence and feeding biology of some nematode spe- settlement dispersal of suspended meiobenthic copepods cies in estuarine aufwuchscommunities. Cah Biol Mar 25: J Mar Res 53509-645 287-303 Cerlach SA (1977) Attraction to decaying organisms as a Buerkel E (1901) Biologische Studien iiber die Fauna der possible cause for patchy distribution of nematodes in a Kieler Fohrde. Lipsius & Tischler. Kiel Bermuda beach. Ophelia 16151-165 Coomans A. De Grisse A (1981) Sensory structures. In: Gray JS (1966a) Factors controlling the localizations of popu- Zuckerman BM, Rohde RA (eds) Plant parasitic nema- lation~of Protodnlus symbioticus Giard. J Anim Ecol 35: todes, Vol 111.Academic Press, New York, p 127-174 435-442 Crisp DJ, Meadows PS (1963) Adsorbed layers: the stimulus to Gray JS (1966b) The attractive factor of intertidal sands to settlement in barnacles. Proc R Soc Lond B 158:364-387 Protodnlus symbioticus Giard. J Mar Biol Assoc UK 46: Croll NA (1970a) The behaviour of nematodes: their activity, 627-645 senses and responses. Edward Arnold, London Gray JS (1967a) Substrate selection by the archiannelid Croll NA (1970b) An analysis of swarming in Panagrellus Protoddus rubropharyngeus Jagersten. Helgolander Wiss redivivus Nematologica 161382-386 Meeresunters 15:253-269 Croll NA, Sukhdeo VK (1981) Hierarchies in nematode Gray JS (196713) Substrate selection by the archlannelid Pro- behavior. In: Zuckerman BM, Rohde RA (eds) Plant todrdus hypoleucos Armenante. J Exp Mar Biol Ecol 1: parasitic nematodes, Vol 111.Academic Press, New York, 47-54 p 227-251 Gray JS (1968) An experimental approach to the ecology of Dietrich G, Kalle K (1957) Allgemeine Meereskunde. Eine the harpacticid Leptastacus constrictus Lang. J Exp Mar Einfiihrung in die Ozeanographie. Gebriider Borntraeger, Biol Ecol2:278-292 Berlin Gray JS, Johnson RM (1970) The bacteria of a sandy beach as Doncaster CC, Webster JM (1968) Clumping of the plant an ecological factor affecting the interstitial gastrotrich parasitic nematode Ditylenchus d~psaciin water. Nemato- Turbanella hyalina Schultze. J Exp Mar Blol Ecol 4: logica 14:131-136 119-133 Dusenbery DB (1973) Countercurrent separation: a new Crewal PS, Wright DJ (1992) Migration of Caenorhabditis method for studying behavior of small aquatic organisms. elegans (Nematoda: Rhabditidae) larvae towards bacteria Proc Natl Acad Sci USA 70:1349-1352 and the nature of the bacterial stimulus. Fundam Appl Dusenbery DB (1974) Analysis of chemotaxis in the nematode Nematol 15:159-166 Caenorhabditis elegans by countercurrent separation Hedgecock EM, Russell RL (1975) Normal and mutant ther- J' EXP 2001 188.41-47 motaxis in the nematode Caenorhabditis elegans. Proc Dusenbery DB (1975) The avoidance of D-tryptophan by Natl Acad Sci USA 72 4061-4065 the nematode Caenorhabditis elegans. J Exp Zool 193: Heip C, Vincx M, Vranken G (1985) The ecology of marine 413-418 nematodes. Oceanogr Mar Biol Annu Rev 23:399-485 Dusenbery DB (1976a) Chemotactic behavior of mutants of Huettel RN (1986) Chemical communicators in nematodes. the nematode Caenorhabditis elegans that are defective J Nematol 18:3-8 in their attraction to sodium chloride. J Exp Zool 198: Jansson HB, Nordbring-Hertz B (1983) The endoparasitic 342-351 nematophagous fungus Mena coniospora infects nema- Dusenbery DB (1976b) Attraction of the nematode Caenor- todes specifically at the chemosensory organs. J Gen habditis elegans to pyridine. Comp Biochem Physiol C Microbiol 130:39-43 Comp Pharmacol53:l-2 J'ensen P (1981) Phytochernical sensitivity and sw~rnming Dusenbery DB (1980) Behavior of free-living nematodes. In: behaviour of the free-living marine nematode Chroma- Zuckerman Bhl (ed) Nematodes as biological models. dorita tenuis. Mar Ecol Prog Ser 4:203-206 Academic Press, New York, p 127-158 Jensen P (1987) Feeding ecology of free-living aquatic nema- Dusenbery DB (1983) Chemotactic behavior of nematodes. todes. Mar Ecol Prog Ser 35:187-196 J Nematol 15:168-173 Kessin RH, Gundersen GG, Zaydfudim V, Grimson M, Dusenbery DB (1988) Avoided temperature leads to the sur- Blanton RL (1996) How cellular slime molds evade nema- face: computer modeling of sllrne mold and nematode todes. Proc Natl Acad Sci USA 93:4857-4861 thermotaxis. Behav Ecol Sociobiol 22:219-223 Klingler J (1965) On the onentation of plant nematodes and of Dusenbery DB (1989) A simple animal can use a complex some other soil animals. Nematologica 11:4-18 stimulus pattern to find a location: nematode thermotaxis Lambshead PJD (1993) Recent developments in marine ben- in soil. Biol Cybern 60:431-438 thic biodiversity. Oceanis 19:5-24 Dusenbery DB (1992) Simultaneous tracking of animals by Lee JJ, Tietjen JH, Mastropaolo C, Rubin H (1977) Food qual- image analysis In. Hader DP (ed) Image analysis in blo- ity and the heterogeneous spatial distribut~on of meio- logy. CRC Press. Boca Raton, FL, p 329-341 fauna. Helgolander Wiss Meeresunters 30:272-282 Dusenbery DB (1996) Information IS where you find it. Biol Lorenzen S, Prein M, Valentin C (1987) h4ass aggregations of Bull 191:124-128 the free-living marine nematode Pontonema vulgare Edmunds JE, Mai MrF (1967) Effect of Fusarium oxysporum on (Oncholaimidae) in organically polluted fjords. Mar Ecol movement of Pratylenchus penetrans toward alfalfa roots. Prog Ser 37:27-34 Phytopathology 57:468-471 Meyers SP, Hopper BE (1966) Attraction of the marine nema- Flndlay SEG (1981) Small-scale spatial distribution of melo- tode, Metoncholaimus sp., to fungal substrates. Bull Mar fauna on a mud- and sandflat. Estuar Coast Shelf Sci 12: Sci 16:142-150 471-484 Meyers SP, Hopper BE (1967) Studies on marlne fungal- Findlay SEG (1982) Influence of sampling scale on apparent nematode associations and plant degradation. Helgolan- distribution of meiofauna on a sandflat. Estuaries 5: der Wiss Meeresunters 15:270-281 322-324 Moens T, Vincx M (1997) Observations on the feeding ecology Fleeger JW, Yund PO. Sun B (1995) Active and passive pro- of estuarine nematodes. J Mar Biol Assoc UK 77:211-227